Vision Research 71 (2012) Contents lists available at SciVerse ScienceDirect. Vision Research. journal homepage:

Transcription

1 Vision Research 71 (2012) Contents lists available at SciVerse ScienceDirect Vision Research journal homepage: Attention is needed for action control: Further evidence from grasping Constanze Hesse a,b,, Thomas Schenk c, Heiner Deubel d a Cognitive Neuroscience Research Unit, Wolfson Research Institute, Durham University, UK b Department of Psychology, University of Aberdeen, UK c Neurology, University of Erlangen Nürnberg, Germany d Allgemeine und Experimentelle Psychologie, Ludwig-Maximilians-Universität, München, Germany article info abstract Article history: Received 13 April 2012 Received in revised form 7 August 2012 Available online 29 August 2012 Keywords: Visual attention Motor control Dual-task Grasping Dorsal Ventral In this study we investigated whether motor and perceptual tasks share common attentional resources. To this end, we used a dual-task paradigm requiring participants to perform grasping movements toward objects of varying size and at the same time to identify a perceptual target presented at the object s location. To ensure that both tasks were performed simultaneously and to prevent participants from adopting a sequential strategy, the perceptual target was always presented after movement onset and could occur at two different moments in time (early vs. late). Our findings show that both, the planning and the control of the movement were altered in the dual-task condition, resulting in prolonged reaction times and delayed adjustment of the grip to object size. Also, the perceptual performance was impaired when both tasks were performed concurrently. These findings are in contrast with previous studies suggesting that only movement planning but not movement control are susceptible to dual-task interferences (Enns & Liu, 2009, chap. 12). Instead, our results give further evidence for the proposition that the dorsal (visuomotor) and ventral (perceptual) stream share the same attentional resources and that attention is required for effective grasping. Ó 2012 Elsevier Ltd. All rights reserved. 1. Introduction Over the last 15 years many studies have provided evidence in favour of the hypothesis that selection-for-perception (ventral) and selection-for-action (dorsal) share a common attentional mechanism (Baldauf & Deubel, 2010; Baldauf, Wolf, & Deubel, 2006; Deubel & Schneider, 2004; Deubel, Schneider, & Paprotta, 1998). The main observation supporting this proposition is that visual discrimination performance is superior when a perceptual stimulus is presented at a location which is also relevant for movement planning. In contrast, when the discrimination stimulus and the movement target are presented at spatially separated locations, visual discrimination performance deteriorates (e.g., Deubel, Schneider, & Paprotta, 1998). While the finding that actionrelevant locations gain more visual attention than locations which are irrelevant for movement planning is relatively uncontroversial, a debate has recently developed on whether dorsal information processing itself (i.e., action planning and control) is constrained by the available attentional capacities (Enns & Liu, 2009, chap. 12; Hesse & Deubel, 2011; Kunde et al., 2007; Liu, Chua, & Enns, 2008). In these studies, the main interest was not in how the Corresponding author. Address: School of Psychology, University of Aberdeen, Kings College, Aberdeen AB24 3FX, UK. address: c.hesse@abdn.ac.uk (C. Hesse). allocation of attention (measured as visual discrimination performance) varies over different spatial locations that are either relevant or irrelevant for movement planning, but on the question of whether movement kinematics are altered when attentional resources have to be shared between a motor and a perceptual task. In a recent study we addressed the issues of whether and how the kinematics of a grasping movement change when a simultaneous perceptual task has to be performed (Hesse & Deubel, 2011). The main result of this study was that the adjustment of the grip to object size was delayed when participants had to perform a simultaneous perceptual task at a different spatial location, indicating that withdrawing processing resources from a movement-relevant location resulted in a less efficient grasp preparation. We interpreted this finding as evidence that motor actions (such as grasping) require attentional capacities and are not completely automated. However, our conclusion conflicts partly with earlier findings. In previous studies it was reported that a concurrent perceptual task only interferes with movement planning (typically characterised by prolonged reaction times (RTs)), but not with the specification of the movement parameters and the on-line control of the action (Enns & Liu, 2009, chap. 12; Liu, Chua, & Enns, 2008). The rationale behind this idea is the assumption that action planning is primarily influenced by the ventral stream whereas action control is considered a function of the dorsal stream (e.g., Glover, 2004; Goodale & Milner, 2004). Recent research seems to suggest that the dorsal and the ventral stream might be controlled /$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.

2 38 C. Hesse et al. / Vision Research 71 (2012) by separated attentional mechanisms (Enns & Liu, 2009, chap. 12; Liu, Chua, & Enns, 2008; Norman, 2002), thus allowing efficient task sharing between perception and action without dual-task costs (but see also, Kunde et al., 2007). For example, Liu, Chua, and Enns (2008) used a dual-task paradigm consisting in the identification of a target letter (ventral task) and in the execution of a pointing movement (dorsal task). They found that while reaction times were prolonged when the perceptual and the motor task had to be performed simultaneously, movement times (MTs) and end-point accuracy were unaffected by the concurrent perceptual task. However, in these tasks participants either started their pointing movement shortly after the perceptual target was presented (Experiment 1), or it was observed that in conditions in which the pointing target could appear before the target letter was presented (Experiment 2), participants preferred to sequence the tasks (waiting with their pointing movement until the perceptual target was presented). Therefore, the lack of dual-task effects on the pointing movements in this study might be primarily related to the fact that the processing resources only had to be shared between tasks at the very beginning of the trial. By the time the hand was in flight toward the pointing target, attentional resources might have been freed up again. This interpretation would also be in line with our previous findings (Hesse & Deubel, 2011) demonstrating that only the early parameters of the grasping movement showed dual-task interferences whereas later parameters (such as maximum grip aperture (MGA) and movement time) remained unaffected by the perceptual task when this task was performed around the time of movement initiation. Thus, to test whether a concurrent perceptual task affects not only movement planning (i.e., RT) but also the specification of movement parameters (such as MGA), it is important to prevent participants from adopting a strategy whereby they can complete the perceptual task before they initiate their motor response. An effective way of preventing this strategy is to make the onset of the perceptual task contingent on the initiation of the motor response and furthermore to vary the onset time for the perceptual target to make its appearance unpredictable. In the current study we adopted both techniques and also measured the movements from start to finish. This allowed us to look for early traces of dual-task costs in the kinematics of the movement. Given the sensorimotor system s capacity to correct in-flight for early errors during later stages of the movement (Hesse & Franz, 2009), early changes may no longer be detectable during the late or final stage of the motor response. This may explain why in Liu, Chua, and Enns (2008) study, where movement kinematics were not measured, interference effects on movement parameters could not be detected. Additionally, the present study was designed to test for an alternative interpretation of our earlier findings (Hesse & Deubel, 2011). As we investigated the effects of splitting attention between two spatial locations, there were always two objects present in the working space (cf. Hesse & Deubel, 2011). Previous research has indicated that the presence of a second object can modify the movement kinematics (e.g., Castiello, 1996). Even though it is still disputed whether the alterations in movement kinematics in the presence of a second object are caused by either: (a) the distractor object attracting attention and therefore evoking a competing motor response which leads to interferences effects (Castiello, 1996, 1998, 1999; Howard & Tipper, 1997; Tipper, Howard, & Jackson, 1997); or (b) by the motor system treating non-target objects as obstacles (Mon-Williams et al., 2001; Tresilian, 1998, 1999), there is convincing evidence that movement kinematics can be altered by the presence of additional objects in the workspace. Although, we tried to address this problem in our previous study by having two objects present in both the dual-task as well as the baseline conditions, one could argue that the two objects were only both relevant during the dual task condition. In the baseline condition, participants could ignore one object and direct their attention to the grasping target only. However, in the dual task condition participants had to pay attention to both objects, since one object contained the targets for the perceptual task and the other one was the target object for the grasping movement. If it were indeed the distracting effect of the second object that explained our previously found interference effects (Hesse & Deubel, 2011), one would expect these interference effects to disappear when only one object is present and when both the perceptual and the motor tasks are directed to the same location. The aim of the present study was therefore to overcome some of the limitations of previous studies investigating dual-task interferences between motor and perceptual tasks. Firstly, by testing whether dual-task interferences also occur when the motor and the perceptual task are performed at the same spatial location (with only one object being presented), we addressed the question of whether changes in movement kinematics are in fact a result of shared attentional resources between tasks or whether the effects can be partly attributed to the presence of additional objects in the workspace. Secondly, by making sure that the perceptual target is only presented after participants have started their movement, we prevented participants from adopting the strategy of task sequencing, and ensured that both tasks are in fact performed simultaneously. 2. Methods 2.1. Participants Eighteen undergraduate and graduate students of Durham University (mean age = 25, age range 18 42) participated in the experiment. All participants were right-handed by self-report and had normal or corrected-to-normal visual acuity. The experiment was approved by the local ethics committee and each participant gave informed consent prior to the study. Participants obtained course credits for their participation, or were paid 5. All participants were naïve with respect to the purpose of the study Apparatus and stimuli The target objects were black plastic rings with four different outer diameters: 45 mm, 50 mm, 55 mm and 60 mm. Each ring had a circular hole in the middle with a constant diameter of 25 mm. All rings were 5 mm in depth and had a small rim on the back at which they could be attached to the setup. The setup consisted of a HANNS.G HX191D LCD monitor (75 Hz) and a Plexiglas frame which was mounted in front of the monitor. The Plexiglas frame was held in place by a custom-built outer frame attached to the monitor (cf. Fig. 1). The Plexiglas frame had a cut-out in the middle which allowed mounting the target rings in front of the monitor leaving the screen visible in the inner cut-out of the rings. The visual stimuli were presented in the inner Fig. 1. Schematic drawing of the experimental apparatus.

3 C. Hesse et al. / Vision Research 71 (2012) cut-out of the rings on the monitor. Before the experiment, the position of the number within the target ring was adjusted such that each participant perceived the number as being in the middle. The visual stimuli consisted of a rapid visual presentation (RSVP) of random digits between 1 and 9. The digits were presented in black on a grey background for 35 ms with a blank interval of 75 ms between each presentation. If the RSVP contained a target digit (in the perceptual baseline and the dual-task conditions), the target digit was presented in white. The size of all digits was 0.6 of visual angle. The size and the presentation duration of the digits in the RSVP were determined in a pilot study (N = 5) adjusting the digits such that participants achieved on average an identification performance of approximately 75%. Participants were seated in front of the monitor with a viewing distance of about 50 cm. A chin rest was used in order to maintain a constant viewing distance. They rested their hand on a starting pin positioned on the table top and vertically aligned to the participants body midline. The distance between the start position of the hand and the centre of the target stimuli was 38 cm. Grasping movements were recorded using an infra-red based Optotrak 3020 system (Northern Digital Incorporation, Waterloo, Ontario, Canada) with a sampling rate of 200 Hz. Infrared lightemitting diodes (IREDs) were placed on the nail of the index finger and the nail of the thumb of each participant. A third IRED was placed on the back of the hand (wrist marker) in order to measure the transport component of the movement. During the experiment, participants wore liquid crystal shutter goggles (PLATO Translucent Technologies, Toronto, Ontario; Milgram, 1987), which rapidly suppress vision by changing from a transparent to an opaque state. Experiments were programmed in MATLAB using the Optotrak Toolbox (Franz, 2004) Procedure Participants sat on an adjustable chair in a well-lit room with their head resting in the chinrest looking straight at the monitor. At the beginning of each trial the shutter glasses turned opaque and the experimenter placed a target ring in the middle of the monitor. Then the experimenter started the trial manually by pressing the key resulting in the opening of the shutter glasses. The experiment consisted of three different tasks which were presented in blocks and counterbalanced across participants. In the perceptual baseline block, the task of the participants was to identify the target digit presented in white during the RSVP. No grasping movement was required. The target digit appeared randomly either after a delay of 350 ms (early) or 700 ms (late) after the opening of the shutter glasses. As soon as the participants had identified the target digit, they reported the number to the experimenter who typed the answer into the computer. After 2 s the shutter glasses closed and the experimenter prepared the next trial. The perceptual block consisted of 32 trials, with each ring size presented 8 times. In every ring size, the target digit occurred either early (350 ms) or late (700 ms) for half of the trials (4 trials per ring size respectively). The experimenter changed the rings after each trial to keep the task perceptually similar to the dualtask. The presentation of all ring sizes and delays was randomised. In the grasping baseline block, the task of the participants was to grasp the target ring with a precision grip (using index finger and thumb) as quickly and accurately as possible immediately after the shutter glasses had opened (a simultaneous auditory go-signal was presented) and to place the ring in front of them on the table. To keep the task perceptually similar to the dual-task situation, the RSVP was presented within the rings but no white target digit was included meaning that the numbers could be ignored. Furthermore, to prevent potential differences between baseline and dualtask conditions occurring due to different fixation strategies in both conditions, we instructed participants to look at the numbers throughout the trial. All ring sizes were presented in random order and each ring size was grasped 8 times resulting in a total of 32 trials. In the dual-task block, participants had to do both tasks simultaneously, that is grasping the target ring whilst reporting the target number presented in the ring. Participants were instructed to do both tasks as accurately as possible. The target digit within the RSVP could either occur early (as soon as the start position was left and the fingers had moved about 1 cm away in z-direction) or late (after half the distance between start position and monitor was reached in z-direction). The two different delays were introduced to prevent participants from predicting the occurrence of the target number and developing certain strategies for the dual-task (e.g., lifting the fingers briefly to cause the target digit to appear followed by the actual grasping movement). Participants grasped the ring and placed it in front of them on the table and reported the target number to the experimenter. Each ring size was grasped 16 times with the target number occurring early in half of the trials. In total the dual-task block consisted of 64 trials with all ring sizes and delays presented randomly. Participants were informed that the target digit would only occur after they had started their movement toward the target Data processing The visual identification performance and the kinematics of the grasping movements measured in the dual-task condition were compared with the performance in the corresponding baseline condition. In order to evaluate the perceptual performance, we calculated the percentage of correctly identified target digits in the perceptual baseline and in the dual-task conditions. In order to test for the effects of the perceptual task on grasping movements, we compared certain kinematic parameters between the grasping baseline and the dual-task conditions. Movement parameters that are usually found to be susceptible to dual-task costs are reaction time (RT) and movement time (MT). Even though the experimental procedure did not differ between the baseline and the dual-task condition until the movement is initiated, the anticipation of an additional task might already interfere with the movement planning process thus delaying movement initiation in the dual-task condition. Movement onset was defined by a velocity criterion (velocities were calculated by differentiating the position signal of the markers). The first frame in which the wrist marker exceeded a velocity threshold of m/s was taken as movement onset. RT is defined as the time between the opening of the shutter glasses (and go-signal) and movement onset. The first frame in which the velocity of the wrist marker dropped under a velocity threshold of 0.05 m/s was taken as the end of the movement. MT was defined as the time between movement onset and end of the movement. Additionally, we calculated the peak velocity (PV) of the wrist during MT, as well as the time at which PV occurred during the movement (TPV) to further characterise the transport component of the grasp. Additionally, we determined several parameters which are known to be related to the accuracy of the grasping movement (for review see, Smeets & Brenner, 1999). Maximum grip aperture (MGA) is defined as the maximum distance in 3D between the markers attached to the thumb and the index finger during MT and linearly related to the size of the target object. The time of occurrence of MGA was calculated as absolute time and as relative time (in % of MT). Finally, we determined how well the aperture was adjusted to the size of the object during the movement. Therefore, we first computed the size of the aperture as mean values binned over 10 samples (40 ms) from movement onset. Then we

4 40 C. Hesse et al. / Vision Research 71 (2012) determined the slope of the function relating object size to aperture size over time by conducting a linear regression analysis at each time point. The slope provides a sensitive measure of how well the grip aperture is adjusted to a specific objects size during the grasp. Data of baseline and the dual-task conditions were analysed using paired samples t-tests and repeated measures analyses of variance (ANOVAs). If the sphericity assumption was violated in the ANOVAs, the degrees of freedom were adjusted using the Greenhouse Geisser correction (Greenhouse & Geisser, 1959). Dependent variables were RT, MT, PV, TPV, MGA, time to MGA (absolute and relative) and the slopes of the function relating object size to the size of the aperture. Values are presented as means + standard errors of the means. If not stated otherwise, a significance level of a =.05 was used for all statistical analyses. 3. Results 3.1. Perceptual performance In this study we were interested in whether the execution of a simultaneous grasping movement affected the perceptual performance in the RSVP, and vice versa. Since a pre-analysis of the data revealed no significant effect of the presentation time of the target digit within the RSVP (early vs. late), neither in the perceptual baseline condition, t(17) = 1.26, p =.23, nor in the dual-task condition, t(17) = 0.62, p =.54, the data was merged over both presentation times. The paired-sample t-test revealed that the identification performance of the participants was significantly better in the baseline condition (77.3 ± 2.5%) than in the dual-task condition (73.8 ± 2.6%), t(17) = 2.73, p =.014. Even though the difference between the baseline condition and the dual-task condition was much smaller than the differences we observed in our previous study in which the perceptual and the grasping task were presented on different spatial locations and no instructions were given about task priority (Hesse & Deubel, 2011), the perceptual performance still decreased significantly when a simultaneous motor task was required. This is a surprising finding, as previous reports seemed to suggest that a concurrent movement to an attended location can result in an enhancement of the perceptual performance (Baldauf & Deubel, 2010) Grasping Dual-task: Effect of target presentation time Before comparing the movement kinematics between the baseline and the dual-task condition, we were interested in whether the time of occurrence of the target digit within the RSVP (early vs. late) affected the movement kinematics in the dual-task condition. To this aim, we merged the data for the dual-task condition over all ring sizes. On average, the target digit occurred at about 80 ± 4 ms after movement onset in the early presentations, and about 290 ± 17 ms after movement onset in the late presentation condition. Paired-sampled t-tests revealed that there was, as expected, no effect of presentation time (early/late) on RT, t(17) = 0.25, p =.81. Similarly, MGA, t(17) = 1.6, p =.12, time to MGA, t(17) = 1.8, p =.09, PV, t(17) = 0.38, p =.71, and TPV, t(17) = 1.2, p =.25 were also unaffected by the presentation time of the target digit. Interestingly however, there was a significant effect on MT, t(17) = 2.3, p =.04, indicating that movements took slightly longer when the target digit occurred later during the movement (early: 687 ± 32 ms; late: 705 ± 37 ms). As the main kinematic landmarks were unaffected by the presentation time of the target, and as we have introduced the variation primarily with the aim to prevent participants from anticipating the occurrence of the target, we merged the data of both presentation times for further analyses Comparison of baseline and dual-task performance In order to test whether the movement kinematics altered when a simultaneous attentional task had to be performed, we compared several kinematic parameters related to the transport and to the grasp component of the movement between the baseline and the dual-task conditions Transport component. The main dependent variable that is known to indicate dual-task costs is RT. On average, RT was about 286 ± 14 ms in the baseline condition and 440 ± 32 ms in the dualtask condition. A 2 (condition: dual-task vs. baseline) 4 (object size: 45 mm, 50 mm, 55 mm, 60 mm) repeated-measures ANOVA confirmed that the difference in RT between the dual-task condition and the baseline condition was significant, F(1,17) = 37.8, p <.001 (Fig. 2). There was no main effect of object size and no interaction effect (both F < 1, p >.44). Thus, already the anticipation that resources have to be shared between a perceptual and a motor task seemed to result in a prolonged movement-planning process. Interestingly, there was no significant prolongation of MTs in the dual-task condition (696 ± 34 ms) compared to the baseline condition (661 ± 24 ms), F(1,17) = 2.75, p =.12 (Fig. 2). Additionally, MT was also unaffected by the variations in object size (p =.28) and there was no significant interaction effect (p =.11). Regarding the peak velocity of the movements, there was a slight tendency for higher PVs for movements performed in the baseline condition (117.9 ± 4.8 cm/s) compared to movements performed in the dual-task condition (111.6 ± 4.1 cm/s). However, the effect did not reach significance, F(1, 17) = 4.08, p =.06. Again, there was no effect of object size and no significant interaction effect (both p >.16). Moreover, the timing of PV was unaffected by any of the experimental variations (all p >.18) Grasp component. As discussed above, we observed in our previous study that a simultaneous attentional task resulted primarily in a delayed adjustment of the aperture to object size (Hesse & Deubel, 2011). A good (but relatively late) indicator for the predictive adjustment of the fingers to object size is MGA. Therefore, we computed MGA and applied a 2 (condition: dual-task vs. baseline) 4 (object size: 45 mm, 50 mm, 55 mm, 60 mm) repeated-measures ANOVA to the data. As expected, there was a Fig. 2. RT, time to MGA (TMGA) and MT in the baseline (grey) and in the dual-task (black) conditions. Values are averaged over all object sizes. Error bars indicate ±1 SEM (between subjects).

5 C. Hesse et al. / Vision Research 71 (2012) significant main effect of object size indicating that MGA was larger for larger objects, F(3, 51) = 289.8, p <.001. There was also a significant effect of condition, F(1, 17) = 5.46, p =.03, revealing that MGA was smaller in the dual-task condition (80.8 ± 1.16 mm) than in the baseline condition (82.6 ± 1.15 mm). A significant interaction effect additionally indicated that the MGA increased more with increasing object size in the baseline condition, F(3, 51) = 10.37, p <.001 (cf. Fig. 3). Regarding the timing of MGA, we replicated our previous finding that MGA occurred later in the dual-task condition (564 ± 36 ms) than in the baseline condition (515 ± 26 ms), F(1,17) = 4.79, p =.04 (Figs. 2 and 4a). There was no main effect of objects size (p =.14) and no interaction (p =.94). The same analysis conducted for the relative time of occurrence of MGA (computed as the percentage of MT) suggests that the grip opening was delayed when the simultaneous attentional task had to be performed (baseline: 77.5% ± 1.5% vs. dual-task: 80.5 ± 1.7%, F(1, 17) = 6.68, p = 0.02). Again there was no main effect of object size (p =.16) and no interaction effect (p =.10). To investigate this issue in more detail, we calculated the size of the aperture in time bins of 10 samples (40 ms) and then calculated the slope of the function relating object size to aperture size using a linear regression analysis. A pre-analysis revealed no significant difference between the slopes in the dual-task conditions dependent on when target was presented (early vs. late, all p >.10). Thus, the data was again merged for further analysis. Fig. 4 shows that the grip opening and the adjustment to the aperture to object size was delayed in the dual-task condition, similarly to our previous findings. The finding that the slopes (relating aperture size to the size of the target object) increased differently over time in the baseline and in the dual-task conditions was statistically confirmed by a repeatedmeasures ANOVA with the factors condition (baseline/dual-task) and time bin revealing a significant interaction effect, F(19, 323) = 8.62, p <.001. Additionally, the analysis revealed a significant main effect of condition, F(1, 17) = 21.67, p <.001, suggesting that the slopes were overall larger in the baseline condition. As expected, there was also a significant main effect of time bin, F(19, 323) = 130.6, p <.001, reflecting the fact that the slopes increase over the course of the movement. Post-hoc tests indicated that all differences between the 6th (220 ms) and 16th (620 ms) time bin were significantly lower than in the baseline condition (we increased our significance level to p <.01 for this analysis, to account for the effects of multiple testing). Overall, these results Fig. 3. Maximum grip aperture as a function of object size in the baseline condition (grey) and in the dual-task condition (black). Error bars indicate ±1 SEM (between subjects). provide evidence that attentional factors play a role in the adjustment of the grasp to object size. 4. Discussion Here, we tested for interference effects between a perceptual identification task and a motor grasping task when both tasks were performed simultaneously and at the same spatial location. The study revealed several important results. Firstly, in line with our previous findings, we observed a delayed opening of the hand as well as a delayed adjustment of the hand to the object s size in the dual-task condition. Thus, even when attention is allocated at the location of the to-be-grasped object (rather than allocated to different locations and to several target objects at the same time) there seem to be significant dual-task costs in grasping. While the alterations in movement kinematics were remarkably similar to those reported when the perceptual and the motor task were performed on separate spatial locations (and in the visual periphery), the dual-task costs in the perceptual task were considerably smaller when both tasks were performed at the same spatial location (cf. Hesse & Deubel, 2011). This discrepancy is, however, not too surprising as it has repeatedly been shown that attention is automatically allocated to locations that are action-relevant (Baldauf & Deubel, 2010; Deubel, Schneider, & Paprotta, 1998). Consequently, one could even hypothesise that the perceptualidentification performance should increase in the dual-task condition of our experiment. That we still observed a small but significant drop of the perceptual performance might be attributed to the fact that the perceptual target was actually presented in the middle of the object and not directly at the locations which were relevant for the grasp (i.e., the contact positions). A recent study indicates that attention is specifically allocated to the contact positions of the grasp, while regions of the objects that are irrelevant for the grasp do not benefit from an increased perceptual performance (Gilster & Deubel, 2011). According to the reasoning, one might claim that even in this experiment the relevant location for grasping and the relevant location for the perceptual target were not strictly identical, as participants were asked to fixate in the middle of the grasping object and not directly at the contact positions. However, recent studies in which eye and hand movement were measured simultaneously suggest that participants naturally prefer to fixate a location which is somewhere in-between the centre of the object and the contact position of the index finger (Brouwer, Franz, & Gegenfurtner, 2009; de Grave et al., 2008). Thus, even if participants are free to move their eyes, they usually do not fixate directly at the anticipated contact positions. This strategy, might be due to the fact that there are usually two contact positions (when grasping with precision grip) and that participants select a location from which they can monitor the approach of both fingers. Hence, we would argue that by presenting the perceptual target in the middle of the to-grasp objects, participants fixated a least a position spatially relatively close to the preferred fixation point on the object. As the distance between the centre of the objects and the outer edges was about 3.43 of visual angle for the largest target object, both contact positions were still well visible when fixating at the centre of the object. Our main finding that the grip opening was delayed in the dualtask condition seems to suggest that there are capacity limitations for dorsal stream tasks. This finding is especially interesting as there is an on-going debate on whether dorsal (visuomotor) and ventral (perceptual) processing share common processing resources (Enns & Liu, 2009, chap. 12; Kunde et al., 2007; Liu, Chua, & Enns, 2008; Norman, 2002). Recently, it was suggested that only action planning (which is considered a function of the ventral

6 42 C. Hesse et al. / Vision Research 71 (2012) Fig. 4. (A) Averaged aperture profiles when grasping target objects of different sizes in the baseline condition (solid lines) and in the dual-task condition (dashed lines) as a function of time. (B) Grip scaling (i.e., the slope of the function relating grip aperture to object size) in the baseline condition (grey) and in the dual-task condition (black), plotted as a function of time. Error bars indicate ±1 SEM (between subjects). stream) relies on the same resources as conscious perceptual tasks whereas action control (executed by the dorsal stream) was considered to be unaffected by an attention-demanding ventral stream task (Enns & Liu, 2009, chap. 12). This interpretation was supported by the finding that RTs (indicator for movement planning), but not MTs (indicator for action control) showed significant dualtask costs when a perceptual and a visuomotor task were performed concurrently (Liu, Chua, & Enns, 2008). However, a possible reason why no dual-task costs were found for the visuomotor task in the study of Liu, Chua, and Enns (2008), might be that they presented the perceptual target either prior to movement initiation (Exp. 1) or shortly afterwards (Exp. 2). Recently, we have shown that when the perceptual task is presented around the time of movement initiation (Hesse & Deubel, 2011), only the early movement characteristics are altered while kinematic parameters characterising the second half of the movement showed no dual-task interference effects. Therefore, one could hypothesise that after participants have identified the target digit, there is enough time to adjust the initially inaccurately planned movement without increasing the overall movement time. The findings from the present study support this hypothesis. Here, the target was always presented after participants had started their movement. In these conditions, we did not only observe a delayed grip opening in the dual-task condition, but also an overall smaller MGA than in the baseline condition. The reduced MGA in the dual-task condition is probably a direct consequence of the delayed adjustment of the grip to object size. In the dual-task condition, participants opened their hand slower and to a similar degree for all objects during the first ms of the movement, resulting in less time to accomplish the wider hand opening required for grasping larger objects after the perceptual target was identified. Consistent with this interpretation is the finding that the differences between the size of MGA in the baseline condition and the dual-task condition were larger for bigger objects than for smaller ones (as indicated by the significant interaction effect, cf. Fig. 3). Alternatively, one could assume that participants increased their movement times in order to compensate for the delayed grip opening during the first half of the movement, and to gain more time for the adjustment of the grasp. In fact, our results indicate that MTs were prolonged in dual-task trials in which the perceptual target occurred later during the movement. The finding that MTs increased when the perceptual target occurred later during the movement but not when the perceptual target was presented shortly after movement initiation may indicate that, as suggested above, movement times only increase when there is not enough time to adjust the grip accurately after the perceptual target is identified and processing resources are freed up. Note, however, that in the present study the different onset times for the perceptual targets were mainly introduced to prevent participants from predicting its occurrence and developing specific strategies. Thus, the question of how perceptual tasks affect movements at different stages was not tested systematically and should therefore be addressed in more detail in further studies before drawing final conclusions. Finally, another interesting finding of our study was that even though participants were aware of the fact that the perceptual target would not occur before they had moved their fingers away from the starting position, RTs were still about 150 ms longer in the dual-task condition. Previous studies investigating the interference effects between perceptual and motor tasks in dual-task condition either presented the perceptual task shortly before the visuomotor task, or introduced a fixed intervals between the tasks (Hesse & Deubel, 2011; Kunde et al., 2007; Liu, Chua, & Enns, 2008). Therefore, one could argue that in the latter case, participants can adopt a strategy of waiting for the perceptual target before initiating their motor response. Results indicated that participants indeed seem to prefer such a sequential strategy (see, Liu, Chua, & Enns, 2008; Exp. 2). Our results show that even when there is no strategic advantage of delaying the start of the movement, dual-tasking is associated with a substantial increase in RTs. This indicates that already the anticipation or the knowledge about the necessity to share resources in an upcoming task may results in less effective movement-planning. In summary, the current study shows that attentional resources are required for effective grasping. Our results suggest that when a concurrent perceptual task has to be performed, participants initiate their grasping movement without accounting for the object s size. The movement program is then adjusted later with the aperture becoming predictive for object size whilst the hand is approaching the target object. It is commonly assumed that the specification of the initial kinematic parameters is a function of the dorsal visual stream (Goodale & Milner, 2004). Our finding that a concurrent perceptual task impairs this initial specification, therefore, suggests two things. First, that dorsal processing is also capacity-limited (see also Kunde et al., 2007) and second, that the dorsal (visuomotor) and ventral (perceptual) stream share the same attentional resources. This second implication is in line with the suggestion that there is a unitary control mechanism of

7 C. Hesse et al. / Vision Research 71 (2012) visual attention that selects the objects for perceptual processing and provides the spatial parameters for goal-directed movements such as reaching and grasping (Schneider, 1995). Acknowledgment Constanze Hesse held a post-doctoral research fellowship of the German Research Council (DFG/HE 6011/1-1) at the time of data collection. References Baldauf, D., & Deubel, H. (2010). Attentional landscapes in reaching and grasping. Vision Research, 50(11), Baldauf, D., Wolf, M., & Deubel, H. (2006). Deployment of visual attention before sequences of goal-directed hand movements. Vision Research, 46, Brouwer, A. M., Franz, V. H., & Gegenfurtner, K. R. (2009). Differences in fixations between grasping and viewing objects. Journal of Vision, 9(1), 24. Castiello, U. (1996). Grasping a fruit: Selection for action. Journal of Experimental Psychology: Human Perception and Performance, 22(3), Castiello, U. (1998). Attentional coding for three-dimensional objects and twodimensional shapes. Experimental Brain Research, 123, Castiello, U. (1999). Mechanisms of selection for the control of hand action. Trends in Cognitive Sciences, 3(7), de Grave, D. D. J., Hesse, C., Brouwer, A. M., & Franz, V. H. (2008). Fixation locations when grasping partly occluded objects. Journal of Vision, 8(7), 11. Deubel, H., & Schneider, W. X. (2004). Attentional selection in sequential manual movements, movements around an obstacle and in grasping. Hove: Psychology Press. Deubel, H., Schneider, W. X., & Paprotta, I. (1998). Selective dorsal and ventral processing: Evidence for a common attentional mechanism in reaching and perception. Visual Cognition, 5, Enns, J. T., & Liu, G. (2009). Attentional limits and freedom in visually guided action. Progress in Brain Research, 176, Franz, V. H. (2004). The Optotrak Toolbox. Retrieved April 15, Available from: < Gilster, R., & Deubel, H. (2011). Divided attention to grasp-relevant locations. Perception, 40(Suppl.), 12. Glover, S. (2004). Separate visual representations in the planning and control of action. Behavioral and Brain Sciences, 27(1), Goodale, M. A., & Milner, A. D. (2004). Plans for action. Behavioral and Brain Sciences, 27(1), Greenhouse, S. W., & Geisser, S. (1959). On methods in the analysis of profile data. Psychometrika, 24(2), Hesse, C., & Deubel, H. (2011). Efficient grasping requires attentional resources. Vision Research, 51(11), Hesse, C., & Franz, V. H. (2009). Corrective processes in grasping after perturbations of object size. Journal of Motor Behavior, 41(3), Howard, L. A., & Tipper, S. P. (1997). Hand deviations away from visual cues: Indirect evidence for inhibition. Experimental Brain Research, 113, Kunde, W., Landgraf, F., Paelecke, M., & Kiesel, A. (2007). Dorsal and ventral processing under dual-task conditions. Psychological Science, 18(2), Liu, G., Chua, R., & Enns, J. T. (2008). Attention for perception and action: Task interference for action planning, but not for online control. Experimental Brain Research, 185(4), Milgram, P. (1987). A spectacle-mounted liquid-crystal tachistoscope. Behavior Research Methods, Instruments, and Computers, 19(5), Mon-Williams, M., Tresilian, J. R., Coppard, V. L., & Carson, R. G. (2001). The effect of obstacle position on reach-to-grasp movements. Experimental Brain Research, 137, Norman, J. (2002). Two visual systems and two theories of perception: An attempt to reconcile the constructivist and ecological approaches. Behavioral and Brain Sciences, 25, Schneider, W. X. (1995). VAM: A neuro-cognitive model for visual attention control of segmentation, object recognition and space-based motor action. Visual Cognition, 2, Smeets, J. B. J., & Brenner, E. (1999). A new view on grasping. Motor Control, 3, Tipper, S. P., Howard, L. A., & Jackson, S. R. (1997). Selective reaching to grasp: Evidence for distractor interference effects. Visual Cognition, 4(1), Tresilian, J. R. (1998). Attention in action or obstruction of movement? A kinematic analysis of avoidance behavior in prehension. Experimental Brain Research, 120, Tresilian, J. R. (1999). Selective attention in reaching: When is an object not a distractor? Trends in Cognitive Sciences, 3,